Predicted Air Conditioning (AC) Prevalence by Core-Based Statistical Area (CBSA) & Census Tract – 2010, 2015, and 2020 Boundaries 

This dataset contains modeled estimates of air conditioning (AC) prevalence across U.S. census tracts using tract boundary definitions for the years 2010, 2015, and 2020. While the spatial boundaries reflect census tract and county definitions for each respective year, the AC prevalence estimates are based on predictive models calibrated using housing characteristics from 2021. This allows the dataset to represent contemporary AC conditions aligned with historical geographic definitions for longitudinal and comparative analysis.

To ensure consistency with known housing stock, the number of predicted units by AC type was adjusted to match the total number of occupied housing units in each census tract, based on 2020 Decennial Census data. Adjustments were applied using proportional scaling. For the 2010 and 2015 datasets, we used crosswalks from the National Historical Geographic Information System (NHGIS) to align 2020 housing unit data with earlier tract boundaries, accounting for geographic changes over time.

In addition, metropolitan status was appended using Core-Based Statistical Area (CBSA) identifiers to help distinguish between metropolitan, micropolitan, and non-core areas.


Adjustments:
- We compared with the occupied housing units from the decennial census in a tract, all predicted counts were proportionally scaled.
- The final dataset includes adjusted counts and percentages to reflect this correction.

Data Fields:
Each file includes the following fields:
- GEOIDCN: The county FIPS code (a 5-digit code combining state and county identifiers).
- GEOID: The 11-digit census tract identifier (state + county + tract).
- CBSAFP: Core-Based Statistical Area (CBSA) FIPS code.
- NAME: Name of the metropolitan or micropolitan area associated with the tract, if applicable.
- ACtype: The type of air conditioning unit (e.g., Central, Window, NoAC).
- Adjusted_Count: The number of housing units in the tract predicted to have the given ACtype, after adjustment.
- Adjusted_Percentage: The percentage of housing units in the tract with the given ACtype, based on adjusted counts. Rounded to two decimal places.

Coordinate Reference System:
Data are based on tract geometries as published by the U.S. Census Bureau for the 2010 boundary definition and use the native CRS from those shapefiles.

File Names:
- summary_metro_adjusted_2010.csv
- summary_metro_adjusted_2015.csv
- summary_metro_adjusted_2020.csv

Each file corresponds to a different set of input housing features (from 2010, 2015, and 2020 respectively), while maintaining the 2010 geographic boundary reference.

Citation:
If you use this dataset, please cite:
Ahn, Y. & Uejio, C. (2025). A Comprehensive Dataset of Residential Air Conditioning Prevalence in the Continental United States. [Data file].

Contact:
Yoonjung Ahn
Department of Geography & Atmospheric Science
University of Kansas
yoonjung.ahn@ku.edu
